It is well known that holography can be used to store massive digital data in a small storage volume, such as a photo-refractive crystal, for a later readout.
A number of configurations exist for holographic storage systems, particularly with respect to the multiplexing techniques used for maximizing storage density. A commonly used configuration is shown in FIG. 1. The primary components of this system are an input device, a holographic storage medium such as a photo-refractive crystal, and an output device. The input device, usually implemented by a spatial light modulator (SLM), consists of an array of amplitude modulators. The output device is a detector array such as a charge-coupled-device (CCD) which receives the output radiation from the storage medium during a hologram readout.
During a hologram-write operation, data are inputted into the SLM in the form of amplitude modulations of the pixels of the SLM. A laser beam illuminates the SLM, and the resulting object beam emanating from the SLM pixels interferes with a write reference beam in the storage medium to create a large number of gratings, one for each SLM pixel. During a hologram-readout operation, a read reference beam illuminates the storage medium, resulting in reflected radiation from each of the stored gratings. The radiation from each grating is detected by a CCD pixel. As an example, for a SLM and a CCD with 1024 by 1024 pixels each, a single hologram (page) stores 1024.times.1024=1 Mbit of binary data per page. If more than two gray levels are used, the data per page increases by the number of bits. Thus, for 16 gray levels (4 bits), the storage density increases by a factor of four, to 4 Mbits per page in the above example.
Noise sources in holographic storage systems have a number of physical origins. The primary sources are (1) light scattered from the storage medium, (2) light scattered from optical components, (3) intra-page pixel crosstalk, (4) inter-page crosstalk and (5) detector noise. Intra-page and inter-page crosstalk can be reduced by conventional design techniques, at the cost of some decrease in storage density. Light scattered from optical components can also be minimized using conventional techniques. However, it is more difficult to reduce the effect of light scattered from the storage medium, called scatter noise, which severely affects the detection signal-to-noise ratio (SNR), a critical parameter directly linked to the performance of holographic storage systems.
When the holographic storage medium is illuminated by the read reference beam, the detectors receive electric fields from both the stored data and from optical scatterers. Scattering from optical components, external to the holographic storage medium, can be minimized using conventional techniques such as anti-reflection coatings. However, scattering from the storage medium itself represents a much more difficult problem since it originates within the same physical volume as the data and since its spectral characteristics are identical to those of the data electric field.
Furthermore, in some storage media such as iron-doped LiNbO.sub.3, scatter noise has been observed to increase with increasing iron doping level. A higher iron concentration is desirable because it improves other system parameters such as hologram erasure time. Thus, reduction of scatter noise will allow other system parameters to be optimized. The invention decouples the effect of scatter noise from such other design considerations, thereby enabling improved overall system performance.
The present invention discloses a method and an apparatus for writing data information into the holographic storage medium such that the detected scatter noise component of the electric field emerging from the holographic storage medium is substantially reduced. Specifically, the invention reduces scatter noise which has the characteristics of laser speckle. The scatter noise intensity is a random complex variable with finite mean. The voltage recorded by each pixel of the CCD array during a hologram readout, due to incident data signal and noise electric fields, can be written in complex variable notation as: EQU V.about..vertline.E.sub.d +E.sub.n e.sup.i.phi. .vertline..sup.2(1)
or EQU V.about.E.sub.d.sup.2 +2E.sub.d E.sub.n cos (.phi.)+E.sub.n.sup.2(2)
where E.sub.d and E.sub.n are the magnitudes of the data and scatter noise electric fields, respectively, and .phi. is their relative phase shift. The proportionality constant between the detected voltage V and the right hand side of Equation (2) depends on detector characteristics and other system parameters, and can be measured or calculated. If the scatter noise is due to a large number of stationary sites in the holographic storage medium, then it has the characteristics of laser speckle. In that case the scatter noise intensity, I.sub.n, which is equal to E.sub.n.sup.2, is a random variable with an exponential probability distribution and the phase, .phi., is uniformally distributed in the interval [-.pi.,.pi.]. The probability density function of the scatter noise intensity I.sub.n is: ##EQU1## where &lt;I.sub.n &gt; denotes the average value of I.sub.n. Although the present invention is described for the case where the scatter noise electric field has the characteristics of speckle, it has general applicability, regardless of the statistics of the scatter field.
The present invention makes use of the fact that, even though the speckle intensity and phase are random variables, the speckle pattern is repeatable from read pulse to read pulse if the interaction geometry and laser beam parameters remain the same.
According to Equation (2), the scatter noise has two contributions, represented by the second and third terms on the right hand side of Equation (2). In a practical system, E.sub.n is much smaller than E.sub.d, and the second term dominates the expression for scatter noise. As an example, for E.sub.n.sup.2 /E.sub.d.sup.2 =10.sup.-2, V.about.1+0.2*cos (.phi.)+0.01. For .phi. near 0 or .pi., the maximum noise voltage is 0.01 without the second term, and approximately .+-.0.2 with the second term included. The corresponding voltage SNR is 100 in the first case and only 5 in the second case.
The present invention discloses a method and an apparatus for eliminating or greatly decreasing the second term by forcing the value of the phase .phi. in Equation (2) to be 90 degrees, thereby increasing the detection signal-to-noise ratio.